Eur Spine J (2003) 12 (Suppl.

2) : S104–S112
DOI 10.1007/s00586-003-0613-0 REVIEW

Mikayel Grigoryan Recognizing and reporting osteoporotic

Ali Guermazi
Frank W. Roemer vertebral fractures
Pierre D. Delmas
Harry K. Genant

Received: 31 July 2003

Abstract Vertebral fractures are the DXA systems delivering “high-reso-
Accepted: 4 August 2003 hallmark of osteoporosis, and occur lution” lateral spine images offers a
Published online: 11 September 2003 with a higher incidence earlier in life potential practical alternative to radio-
© Springer-Verlag 2003 than any other type of osteoporotic graphs for clinical vertebral fracture
fractures. It has been shown that both analysis. The advantages of using
symptomatic and asymptomatic ver- DXA over conventional radiographic
tebral fractures are associated with devices are its minimal radiation ex-
increased morbidity and mortality. posure and high-speed image acqui-
Morbidity associated with these frac- sition. It also allows combined eval-
tures includes decreased physical uation of vertebral fracture status and
function and social isolation, which bone mass density, which could be-
have a significant impact on the pa- come a standard for patient evalua-
tient’s overall quality of life. Since tion in osteoporosis. The disadvan-
the majority of vertebral fractures do tage of DXA use is that upper tho-
not come to clinical attention, radio- racic vertebrae cannot be evaluated
graphic diagnosis is considered to be in a substantial number of patients
the best way to identify and confirm due to poor imaging quality. We
the presence of osteoporotic verte- truly believe that the that there is a
bral fractures in clinical practice. major role for radiologists and clini-
Traditionally, conventional lateral ra- cians alike to carefully assess and di-
diographs of the thoracolumbar spine agnose vertebral fractures using stan-
have been visually evaluated by radi- dardized grading schemes such as
ologists or clinicians to identify ver- the one outlined in this review.
tebral fractures. The two most widely Quantitative morphometry is useful
M. Grigoryan · A. Guermazi · used methods to determine the sever- in the context of epidemiological
F. W. Roemer · H. K. Genant (✉) ity of such fractures in clinical re- studies and clinical drug trials; how-
Osteoporosis and search are the semiquantitative as- ever, the studies would be flawed if
Arthritis Research Group,
Department of Radiology, sessment of vertebral deformities, quantitative morphometry were to be
University of California San Francisco, which is based on visual evaluation, performed in isolation without addi-
350 Parnassus Avenue, and the quantitative approach, which tional adjudication by a trained and
San Francisco, CA 94117, USA is based on different morphometric highly experienced radiologist or
Tel.: +1-415-4763680,
Fax: +1-415-4768550,
criteria. In our practice for osteo- clinician.
e-mail: Harry.Genant@oarg.ucsf.edu porosis evaluation we use the Genant
semiquantitative approach: an accu- Keywords Osteoporosis · Vertebral
P. D. Delmas
Hopital Edouard Herriot, rate and reproducible method tested fractures · Semiquantitative
Place d’Arsonval, Pavillion F, and applied in many clinical studies. assessment · Bone mineral density ·
69347 Lyon, Cedex 03, France The newest generation of fan-beam Quantitative morphometry
 S105

Introduction cal symptoms such as pain or height loss. Therefore the

evaluation of spinal radiographs for prevalent and inci-
Osteoporosis is a serious public health problem. The inci- dent vertebral fractures is important in both clinical and
dence of osteoporotic fractures increases with age. As life epidemiological evaluation of patients with established osteo-
expectancy increases for a greater proportion of the world’s porosis and populations at risk for developing it. Fewer
population, the financial and human costs associated with than 1% of back pain episodes are related to vertebral
osteoporotic fractures will multiply exponentially. Ac- fractures [10]. Therefore vertebral fractures are often not
cording to the International Osteoporosis Foundation, more suspected in patients reporting back pain, unless associ-
than 40% of middle-aged women in Europe will suffer ated with trauma. Trauma-related fractures, however, are
one or more osteoporotic fracture during their remaining not considered as classical osteoporotic fractures. Histori-
lifetime [23]. cal height loss is also difficult to assess clinically. Some
Vertebral fractures are the hallmark of osteoporosis height loss is expected with aging due to compression of
and occur with a higher incidence earlier in life than any the intervertebral discs and postural changes. However,
other type of osteoporotic fractures, including hip frac- height loss could also be due to multiple fractures, which
tures [34]. The importance of fragility fractures, of which represent significant and irreparable damage. Therefore it
vertebral fractures are the most common, was acknowl- has been concluded that height loss is an unreliable indi-
edged by the World Health Organization classification cri- cator of fracture status until it exceeds 4 cm [9]. As a result
teria for osteoporosis evaluation [51]. The criterion of the vertebral fractures are often not being considered in clini-
World Health Organization defines “severe osteoporosis” cal patient evaluation, and it is relatively uncommon for
as “low bone mass (T score below –2.5) in the presence of patients to be referred for radiographs in the course of os-
one or more fragility fractures.” teoporosis testing. Improvements in detecting and report-
The definition of osteoporosis is centered on the level ing vertebral fractures in patients with osteoporosis would
of bone mass, which is measured as bone mineral density increase the potential of therapeutic intervention to pre-
(BMD). BMD measurements are widely used to estimate vent subsequent fractures.
the risk of osteoporotic fractures and individuals who are
at risk for osteoporotic fractures are usually referred for
BMD measurements under the current standard of care. In Radiographic assessment of vertebral fractures
addition, many other risk factors have been identified, some
of which are known to add to the risk independently of Radiographic diagnosis is considered to be the best way to
BMD measurements. The combination of BMD with such identify and confirm the presence of osteoporotic verte-
risk factors increases the gradient of risk/standard devia- bral fractures in clinical practice. Traditionally, conven-
tion than that achieved by BMD alone. Several clinical tional lateral radiographs of the thoracolumbar spine have
trials have demonstrated that a substantial improvement in been visually evaluated by radiologists or clinicians to
the assessment of the risk for future fractures can be ac- identify vertebral fractures. However, there is still no in-
complished by the assessment of prevalent vertebral frac- ternationally agreed definition for vertebral fracture. One
tures in combination with BMD measurements [2, 5, 15, global prospective study (the IMPACT study [6]) compared
27, 31, 36, 39, 41]. Nonetheless, it remains a common the results of local radiographic reports from five conti-
clinical practice to consider “low” BMD to be a risk fac- nents with that of subsequent central readings in more
tor irrespective of the presence of vertebral fractures. than 2,000 postmenopausal women with osteoporosis. This
study demonstrated that vertebral fractures were frequently
underdiagnosed radiologically worldwide, with false-neg-
Clinical identification of vertebral fractures ative rates as high as 30% despite a strict radiographic
protocol that provided an unambiguous vertebral fracture
It has been shown that both symptomatic and asympto- definition and minimized the influence of inadequate film
matic vertebral fractures are associated with increased quality. It was concluded that the failure was a global
morbidity [9] and mortality [8, 22, 35]. Morbidity associ- problem attributable to either lack of radiographic detec-
ated with these fractures includes decreased physical func- tion or use of ambiguous terminology in reports. There-
tion and social isolation, which have a significant impact fore it is very important to use standardized methods for
on the patient’s overall quality of life [16]. Still, it remains the visual assessment of vertebral fractures.
difficult to determine the exact incidence of osteoporotic Several standardized approaches to describe vertebral
vertebral fractures that occur annually, as a substantial pro- fractures have been proposed. They may serve to facilitate
portion remains clinically undetected. Large-scale prospec- the diagnosis of osteoporosis and to assess the severity or
tive studies demonstrate that only about one of four verte- progression of the disease as well as to rule out nonfrac-
bral fractures becomes clinically recognized [7]. This is ture deformities or normal variants.
due to both the absence of specific symptoms in some and
the difficulty in determining the cause of possible physi-
S106

Visual semiquantitative methods A vertebral deformity (graded 1–3) is present when ha, hm,
of vertebral fracture assessment or hp is reduced by at least 4 mm or 15%. This score, as
with Meunier’s radiological vertebral index, still relies
The first standardized approach was introduced by Smith very much on the type of deformity, i.e., the vertebral shape,
et al. [44] in 1960. They introduced a classification of ver- and there would have to be changes in vertebral shape in
tebral deformities as diagnosed from lateral thoracolum- order to account for incident vertebral fractures on follow-
bar radiographs for the purpose of diagnosing the severity up radiographs. Furthermore, the majority of vertebral
of osteoporosis. This method grades only the most se- fractures consist of a combination of wedge and endplate
verely deformed vertebra on the radiograph. In 1968 Me- deformities, and less frequently posterior deformities. There-
unier [33] proposed an approach in which each vertebra is fore an examiner’s distinction among these deformities is
graded according to its shape or deformity. Grade 1 is as- often quite subjective.
signed to a normal vertebra that has no deformity, grade 2 A vertebral deformity does not always represent a ver-
to a biconcave vertebra, and grade 4 to an endplate frac- tebral fracture, but a vertebral fracture is always a verte-
ture or a wedged or crushed vertebra. Using this approach bral deformity. From a radiological prospective, there are
vertebral bodies T3 (or T7) to L4 are evaluated. A “radio- many potential differential diagnoses for vertebral defor-
logical vertebral index” can be calculated as the sum of mities – osteoporotic fracture, posttraumatic deformity, de-
the grades of all vertebrae, or as the quotient of this sum generative remodeling, Scheuermann’s disease (juvenile
and the number of the vertebrae. kyphosis), congenital anomaly, neoplastic deformity, and
Kleerkoper et al. [26] modified Meunier’s radiological Paget’s disease – and the correct qualitative classification
vertebral index and introduced the so-called “vertebral de- of vertebral deformities can be accomplished only by vi-
formity score.” In the vertebral deformity score each ver- sual inspection and expert interpretation of the radiograph.
tebra from T4 to L5 is assigned an individual score from This perspective on vertebral fracture diagnosis is perhaps
0 to 3 depending on the type of deformity. This grading reflected at its best in the semiquantitative fracture assess-
scheme is based on the reduction in the anterior, middle, ment method proposed by Genant et al. [12, 13, 14, 15]
and posterior vertebral heights (ha, hm, and hp, respectively). This method provides an insight into the severity of a

Fig. 1 Schematic diagram of

semiquantitative grading scale
for vertebral fractures. (From
Genant et al. [13])
 S107

fracture which is assessed solely by visual estimation of mation but cannot be clearly assigned to grade 1 fractures is
the extent of a vertebral height reduction and morpholog- sometimes also utilized. In addition to height reductions,
ical change, and vertebral fractures are differentiated from careful attention is given to alterations in the shape and
other, nonfracture deformities. In Genant’s visual semi- configuration of the vertebrae relative to adjacent verte-
quantitative assessment (Fig. 1) each vertebra receives a brae and expected normal appearances. These features add
severity grade based upon the visually apparent degree of a strong qualitative aspect to the interpretation. For exam-
vertebral height loss. Unlike the other approaches the type ple, vertebral deformities due to degenerative changes
of the deformity (wedge, biconcavity, or compression) is should be ruled out, whereas an endplate vertebral frac-
no longer linked to the grading of a fracture in this ap- ture can be identified without a 20% reduction in the ver-
proach. tebral height. Nevertheless, in experienced, highly trained
Thoracic and lumbar vertebrae from T4 to L4 are graded hands, it makes the approach both sensitive and specific.
on visual inspection and without direct vertebral measure- A “spinal fracture index”) can be calculated from this
ment as normal (grade 0), mildly deformed (grade 1: re- semiquantitative assessment as the sum of all grades as-
duction of 20–25% of height and 10–20% of projected signed to the vertebrae divided by the number of the eval-
vertebral area), moderately deformed (grade 2: reduction uated vertebrae.
of 26–40% of height and 21–40% of projected vertebral An advantage of this semiquantitative approach over
area), and severely deformed (grade 3: reduction of >40% other standardized visual approaches is that the severity of
of height and projected vertebral area; Fig. 2). A grade 0.5 the deformation as the reduction in vertebral height means
designates “borderline” vertebrae that show some defor- can be assessed from serial films and is especially useful
for the interpretation of incident fractures. It considers the
continuous character of vertebral fractures and makes a
meaningful interpretation of follow-up radiographs possi-
ble. Furthermore, inevitably arbitrary decisions regarding
wedge, endplate, or crush deformities, as assessed in some
grading schemes, are not necessary since most fractures

Fig. 2 Lateral thoracic radiograph shows a grade 3 fracture T8 Fig. 3 Degenerative remodeling in middle-thoracic region simu-
and grade 2 fractures of T9 and T11 lating wedge deformities
S108

contain a combination of these features, influenced by the

local biomechanics of the spinal level.
The Genant’s semiquantitative method has been tested
and applied in a number of clinical drug trials and epi-
demiological studies [15, 20, 47, 50, 52]. The repro-
ducibility of the method for the diagnosis of prevalent and
incident vertebral fractures was found to be high, with in-
traobserver agreement of 93–99% and interobserver agree-
ment of 90–99%. This indicates that close agreement among
readers can be reached using this standardized visual semi-
quantitative grading method, and that subjectivity in the
readings can be reduced. This accounts for experienced
and relatively inexperienced readers with reasonable re-
sults.
There are limitations of this semiquantitative grading
scheme that may also apply to other standardized ap-
proaches. For example, from the morphometric data on
normal subjects we know that vertebrae in the middle tho-
racic spine (especially in women) and thoracolumbar junc-
tion (especially in men) are slightly more wedged than in
other regions (Fig. 3) [3, 30, 32, 40]. As a result these nor-
mal variations may be misinterpreted as mild vertebral de-
formities, thereby falsely increasing prevalence values for
vertebral fractures. The same applies to a lesser extent to Fig. 4 Example of six-point placement in quantitative vertebral
morphometry
the middle to lower lumbar spine, where some degree of
biconcavity is frequently observed [26, 45]. Accurate di-
agnosis of prevalent fractures which requires distinguish- ition [3, 42]. Typically six points are used to derive the an-
ing between normal variations and the degenerative changes terior height (ha), the central (middle or middle-vertebral,
from true fractures still depends on the experience of the hm) height, and the posterior height (hp; Fig. 4). This ex-
observer. It has been argued that the diagnosis of mild ver- clusively quantitative approach has, however, a number of
tebral fractures (grade 0.5–1) in particular may be quite drawbacks including projectional effects that significantly
subjective, and that these fractures may be unrelated to influence the reliability of these measures performed in
osteoporosis [45]. However, mild fractures are also asso- isolation.
ciated with a lower bone density and to a certain extend In general, a substantial number of mild deformities de-
predict future vertebral fractures [1]. tected by visual reading are missed by the quantitative tech-
Other limitations may apply for the diagnosis of inci- nique when applying the common threshold values for re-
dent fractures. The reader may sometimes feel that even duction in vertebral heights such as 15–20% or 3 SD de-
though a further height reduction is seen in a previous ver- crease. Furthermore, a significant number of false positives
tebral fracture, it may not be justified to assign a higher are found with quantitative techniques. The choice of point
fracture grade on a serial radiograph, since some degree of placement in the quantitative technique, but especially the
settling or remodeling generally occurs. Therefore in gen- choice of the threshold for defining vertebral deformity,
eral, serial radiographs including the baseline radiograph gives results that vary in specificity and sensitivity. Most
of a patient should be viewed together so that incident of the moderate to severe deformities are detected by both
fractures can be readily identified as only those progres- techniques. However, only expert visual evaluation can
sive changes that lead to a full increase in deformity grade detect mild and subtle deformities, as well as appreciate
or from a questionable deformity (grade 0.5) to a definite anatomical, pathological and technical issues that bear on
fracture. the evaluation of fracture detection.
The strength of a semiquantitative approach is that it
makes use of the entire spectrum of visible features that
Quantitative morphometry and its comparison are helpful in identifying deformities [15, 49]. The visual
with the semiquantitative methods interpretation, when performed by the expert eye, also
separates true deformities from normal or anomalous ver-
Quantitative morphometric assessment of vertebral defor- tebrae. In addition to changes in dimension, vertebral de-
mity was introduced in order to obtain an objective and formities are generally detected visually by the presence
reproducible measurement, using rigorously defined point of endplate deformities, the lack of parallelism of the end-
placement and well-defined algorithms for fracture defin- plates, and the general altered appearance compared with
 S109

neighboring vertebrae. Some of these visual characteris- on Vertebral Fractures suggested the following procedural
tics are not captured by the six-digitization points used in requirements for a qualitative (semiquantitative) assess-
quantitative techniques; this can cause some deformities ment of vertebral fractures in osteoporosis research [25]:
to remain undetected. For example, only an experienced
– Assessments should be performed by a radiologist or
observer can make the subtle distinctions between a frac-
trained clinician who has specific expertise in the radi-
tured endplate and wedge shaped appearance caused by
ology of osteoporosis.
the remodeling of the vertebral bodies in degenerative disc
– Qualitative and semiquantitative assessments should be
disease (Fig. 3). This is often interpreted as a wedge frac-
performed according to a written protocol of fracture
ture in quantitative studies.
definitions, which are sufficiently detailed that the read-
In the absence of distinct characteristics of a fracture,
ings can be reproduced by other experts. Reference to
however, a reader using a visual approach could rather ar-
an atlas of standard films or illustrations may be help-
bitrarily consider a mild wedge deformity normal, anom-
ful. It is recommended that a standardized protocol be
alous, or fractured; in such a case, a well-defined quanti-
developed by a consensus of expert radiologists.
tative criterion could be useful. Even here, however, with
– The definition of fracture should include deformities of
borderline wedge deformity, small subjective difference
the endplates and anterior borders of vertebral bodies,
in joint placement could result in considerable variation in
as well as generalized collapse of a vertebral body.
fracture/nonfracture discrimination of sequential films or
– Grading of the extent of each fracture should employ
even on the same film.
discrete, mutually exclusive categories. An atlas of stan-
Most incident fractures, as with prevalent fractures, are
dard films and illustrations may again help to assure
easily identifiable visually on sequential radiographs. The
consistency.
unavoidable variation in position and parallax may result
in differences in point placement on follow-up radio- There is some subjectivity in each method, and perform-
graphs. This can result in the morphometric detection of ing the grading in discrete, exclusive categories may be
an incident fracture that would be interpreted visually as problematic at times, particularly for prevalent fractures.
simply an alteration in projection. These sources of false- However, for the assessment of vertebral fractures in the
positive or false-negative interpretation are especially com- form of a fracture/nonfracture dichotomy, trained readers
mon when parallax problems due to radiographic technique have achieved excellent results. After all, the fracture/
or patient positioning are encountered. nonfracture distinction may be the most important, and
Intraobserver variability for a semi-quantitative ap- the semiquantitative standardized grading schemes may
proach depends on experience and training. The same be the instruments to make this diagnosis reliable and valid.
however, is true for digitizing techniques: an experienced Ensuring the reliability of the interpretation of incident
observer is more consistent in the placement of the points vertebral fractures on serial radiographs requires close at-
for digitization. tention to the procedure. Serial radiographs of a patient
A number of comparative studies have evaluated the should always be viewed together in chronological order
relative performance of the quantitative morphometric and to accomplish a thorough and reliable analysis of all new
the semiquantitative methods and moderate correlations fractures. Because a vertebral fracture is a permanent event
were found in most of them [1, 17, 29, 52]. The concor- that is unlikely to vanish on follow-up radiograph, tempo-
dance was high for fractures defined as moderate or se- ral blinding does not appear to be any use: most readers
vere by semiquantitative reading. There was, however, a easily identify a temporal sequence of films by new de-
significant discordance for fractures defined as mild in the formities as well as by progressive disc degeneration and
semiquantitative reading. Additionally, the interobserver osteophyte formation, which are universal among the el-
agreement was demonstrated to be better for the visual derly.
semiquantitative approach. The authors of these studies
concluded that quantitative morphometry should not be
performed in isolation, particularly when applying highly Alternatives to radiographic assessment
sensitive morphometric criteria at low threshold levels of vertebral fractures
without visual assessment to confirm the detected preva-
lent or incident vertebral deformities as probable fractures. Because of the difficulty in identifying vertebral fractures
clinically, and the practical difficulties preventing routine
radiographic assessment at the point of care, vertebral
Standardization of visual approaches fracture status is frequently unknown at the time of patient
to vertebral fracture assessment evaluation for BMD [18]. Hence the interest in morpho-
metric assessment from dual X-ray absorptiometry (DXA)
In an effort to develop a standardized consensus protocol images was a natural consequence of the need for quanti-
for the visual assessment of vertebral fractures, the United tative fracture evaluation in pharmaceutical trials. The
States National Osteoporosis Foundation’s Working Group main advantage of the morphometric X-ray absorptiome-
S110

than dual energy scan modes due to substantially lower

signal to noise in the images and can be performed during
suspended respiration. High-dose, dual-energy acquisi-
tions, while slower, generally provide higher bone con-
trast images and sometimes reduce artifacts.
The use of fan-beam DXA images for quantitative
(morphometric) assessment of spinal fractures has been
reported in both research applications and pharmaceutical
trials [4, 11, 19, 21, 28, 37, 38, 46]. Clinical studies
demonstrated the feasibility of visual evaluation of fan-
beam lateral DXA spine images compared to conventional
lateral spine radiographs in postmenopausal women, with
a strong overall agreement of 96.3% [37, 38]. This agree-
ment was approximately as strong as that found among
different morphometric techniques [15, 21]. The images
permitted visual assessment of about 90% of all vertebrae.
The main shortcoming of the MXA scans in comparison
with conventional radiographs is the inferior image qual-
ity that limits the evaluation of vertebrae in the upper tho-
Fig. 5 Rapid (10-s) “high-resolution” fan-beam DXA imaging al- racic spine. This is less of a concern if MXA is used as a
lows both visual (a) and quantitative (b) assessment of vertebral
fractures screening tool for conventional radiography and this ap-
proach may help reduce the radiation dose in the diagno-
sis and monitoring of osteoporosis.
try technique is that the radiation dose to the patient is
substantially reduced compared with conventional radiog-
raphy. The use of “high-resolution” lateral spine images, Conclusion
obtained with fan-beam X-ray bone densitometry systems
(Fig. 5), offers a potential practical alternative to radio- Vertebral fractures are the most common type of osteo-
graphs for clinical vertebral fracture analysis. “High-reso- porotic fracture, occurring in a substantial portion of the
lution” fan-beam DXA systems, utilizing technology simi- elderly population. Most new vertebral fractures, even
lar to that used by computed tomography (CT) systems, painful ones, remain unrecognized by patients and their
can image the lateral spine in as little as 10 s. In fact CT physicians. It is established that the presence of a verte-
scout scans, with about the same image resolution as fan- bral fracture is a strong risk factor for subsequent osteo-
beam DXA scans, have been used for vertebral fracture porotic fractures, and that those with low bone density and
identification [24, 43, 48]. vertebral fractures are at highest risk. Large-scale clinical
As with radiographs, however, CT images are expen- trials have demonstrated that osteoporosis therapies can
sive and are not available clinically without referral. Con- reverse bone loss and reduce fracture rates, and that these
sequently CT is not generally an option unless performed benefits are most pronounced in patients with low BMD
in conjunction with quantitative CT for BMD assessment. and vertebral fractures. Clinical guidelines promulgated
In contrast, DXA images can be performed at the point of by the National Osteoporosis Foundation, International
care, in conjunction with standard BMD determination, Osteoporosis Foundation, and others recognize the impor-
with a radiation dose as much as 100 times lower than that tance of vertebral fractures, along with BMD, as the key
of conventional radiographs. The most notable strength of risk factors for use in patient evaluation. However, while
radiographs, of course, is image resolution, which is supe- BMD is widely used in patient evaluation, radiological as-
rior to that of DXA images. sessment of vertebral fractures is commonly not performed,
DXA images provide several advantages. The digital or if performed, is inadequately standardized and inter-
nature allows for electronic data storage, digital image en- preted. By understanding the clinical principles of osteo-
hancement and processing, as with magnification and con- porosis diagnosis and management provided in this docu-
trast adjustment, which is not possible with conventional ment and by adopting the radiological guidelines for as-
radiographic techniques. Cone-beam distortion, inherent sessing vertebral fractures provided herein, clinicians world-
in the radiographic technique, is not present when using wide can contribute substantially to reducing the conse-
the scanning fan-beam geometry of DXA devices. Low- quences of this important disease.
dose, single-energy acquisition modes are substantially faster
 S111

References

1. Black D, Palermo L, Nevitt MC, 11. Ferrar L, Jiang G, Eastell R, Peel NF 22. Kado DM, Duong T, Stone KL, Ensrud
Genant HK, Epstein R, San Valentin (2003) Visual identification of verte- KE, Nevitt MC, Greendale GA, Cum-
R, Cummings SR, and the Study of Os- bral fractures in osteoporosis using mings SR (2003) Incident vertebral
teoporotic Fractures Research Group morphometric X-ray absorptiometry. fractures and mortality in older women:
(1995) Comparison of methods for J Bone Miner Res 18:933–938 a prospective study. Osteoporos Int 14:
defining prevalent vertebral deformi- 12. Genant HK, Jergas M (2003) Assess- 589–594
ties: the study of osteoporotic fractures. ment of prevalent and incident verte- 23. Kanis JA, Delmas P, Burckhardt P,
J Bone Miner Res 10:890–902 bral fractures in osteoporosis research. Cooper C, Torgerson D (1997) Guide-
2. Black DM, Arden NK, Palermo L, Osteoporos Int 14 Suppl 3:S43–S55 lines for diagnosis and management of
Pearson J, Cummings SR (1999) 13. Genant HK, Jergas M, Palermo L, osteoporosis. The European Founda-
Prevalent vertebral deformities predict Nevitt MC, Valentin RS, Black D, tion for Osteoporosis and Bone Dis-
hip fractures and new vertebral defor- Cummings SR (1996) Comparison of ease. Osteoporos Int 7:390–406
mities but not wrist fractures. Study of semiquantitative visual and quantita- 24. Katragadda CS, Fogel SR, Cohen G,
Osteoporotic Fractures Research tive morphometric assessment of Wagner LK, Morgan C 3rd, Handel
Group. J Bone Miner Res 14:821–828 prevalent and incident vertebral frac- SF, Amtey SR, Lester RG (1979) Digi-
3. Black DM, Cummings SR, Stone K, tures in osteoporosis. The Study of Os- tal radiography using a computed to-
Hudes E, Palermo L, Steiger P (1991) teoporotic Fractures Research Group. mographic instrument. Radiology 133:
A new approach to defining normal J Bone Miner Res 11:984–996 83–87
vertebral dimensions. J Bone Miner 14. Genant HK, van Kuijk C, Jergas M 25. Kiel D (1995) Assessing vertebral frac-
Res 6:883–892 (1995) Vertebral fracture in osteoporo- tures. National Osteoporosis Founda-
4. Blake GM, Rea JA, Fogelman I (1997) sis. Radiology Research and Education tion Working Group on Vertebral Frac-
Vertebral morphometry studies using Foundation, San Francisco tures. J Bone Miner Res 10:518–523
dual-energy X-ray absorptiometry. 15. Genant HK, Wu CY, van Kuijk C, 26. Kleerekoper M, Nelson DA (1992)
Semin Nucl Med 27:276–290 Nevitt MC (1993) Vertebral fracture Vertebral fracture or vertebral defor-
5. Davis JW, Grove JS, Wasnich RD, assessment using a semiquantitative mity. Calcif Tissue Int 50:5–6
Ross PD (1999) Spatial relationships technique. J Bone Miner Res 8:1137– 27. Kotowicz MA, Melton LJ 3rd, Cooper
between prevalent and incident spine 1148 C, Atkinson EJ, O’Fallon WM, Riggs
fractures. Bone 24:261–264 16. Gold DT (2001) The nonskeletal con- BL (1994) Risk of hip fracture in
6. Delmas PD, Watts N, Eastell R, von sequences of osteoporotic fractures. women with vertebral fracture. J Bone
Ingersleben G, van de Langerijt L, Ca- Psychologic and social outcomes. Miner Res 9:599–605
hall DL (2001) Underdiagnosis of ver- Rheum Dis Clin North Am 27:255– 28. Lang T, Takada M, Gee R, Wu C, Li J,
tebral fractures is a worldwide prob- 262 Hayashi-Clark C, Schoen S, March V,
lem: The IMPACT Study. J Bone 17. Grados F, Roux C, de Vernejoul MC, Genant HK (1997) A preliminary eval-
Miner Res 16 Suppl. 1:S139 Utard G, Sebert JL, Fardellone P uation of the lunar expert-XL for bone
7. Ensrud KE, Nevitt MC, Palermo L, (2001) Comparison of four morphome- densitometry and vertebral morphome-
Cauley JA, Griffith JM, Genant HK, tric definitions and a semiquantitative try. J Bone Miner Res 12:136–143
Black DM (1999) What proportion of consensus reading for assessing preva- 29. Leidig-Bruckner G, Genant HK, Minne
incident morphometric vertebral frac- lent vertebral fractures. Osteoporos Int HW, Storm T, Thamsborg G, Bruckner
tures are clinically diagnosed and vice 12:716–722 T, Sauer P, Schilling T, Soerensen OH,
versa? J Bone Miner Res 14:S138 18. Guermazi A, Mohr A, Grigorian M, Ziegler R (1994) Comparison of a semi-
8. Ensrud KE, Thompson DE, Cauley JA, Taouli B, Genant HK (2002) Identifi- quantitative and a quantitative method
Nevitt MC, Kado DM, Hochberg MC, cation of vertebral fractures in osteo- for assessing vertebral fractures in os-
Santora AC 2nd, Black DM (2000) porosis. Semin Musculoskelet Radiol teoporosis. Osteoporos Int 4:154–161
Prevalent vertebral deformities predict 6:241–252 30. McCloskey EV, Spector TD, Eyres
mortality and hospitalization in older 19. Hans D, Baiada A, Dubouef F, Vignot KS, Fern ED, O’Rourke N, Vasikaran
women with low bone mass. Fracture E, Bochu M, Meunier P. Expert-XL S, Kanis JA (1993) The assessment of
Intervention Trial Research Group. (1996) Clinical evaluation of a new vertebral deformity: a method for use
J Am Geriatr Soc 48:241–249 morphometric technique on 21 patients in population studies and clinical trials.
9. Ettinger B, Black DM, Nevitt MC, with vertebral fracture. Osteoporos Int Osteoporos Int 3:138–147
Rundle AC, Cauley JA, Cummings SR, 6:79 31. Melton LJ III, Kan SH, Frye MA,
Genant HK (1992) Contribution of ver- 20. Heuck AF, Block J, Glueer CC, Steiger Wahner HW, O’Fallon WM, Riggs BL
tebral deformities to chronic back pain P, Genant HK (1989) Mild versus defi- (1989) Epidemiology of vertebral frac-
and disability. The Study of Osteo- nite osteoporosis: comparison of bone tures in women. Am J Epidemiol 129:
porotic Fractures Research Group. densitometry techniques using different 1000–1011
J Bone Miner Res 7:449–456 statistical models. J Bone Miner Res 4: 32. Melton LJ 3rd, Atkinson EJ, Cooper C,
10. Ettinger B, Cooper C (1995) Clinical 891–900 O’Fallon WM, Riggs BL (1999) Verte-
assessment of osteoporotic vertebral 21. Jergas M, Lang TF, Fuerst T (1995) bral fractures predict subsequent frac-
fractures. In: Genant HK, Jergas M, Morphometric X-ray absorptiometry. tures. Osteoporos Int 10:214–221
van Kuijk C (ed) Vertebral fracture in In: Genant HK, Jergas M, van Kuijk C 33. Meunier P (1968) La dynamique du re-
osteoporosis. Radiology Research and (ed) Vertebral fracture in osteoporosis. maniement osseux humain, etudiee par
Education Foundation, San Francisco, Radiology Research and Education lecture quantitative de la biopsie os-
pp 15–20 Foundation Publishers, San Francisco, seuse. Lyon
pp 331–348
S112

34. Meunier PJ, Delmas PD, Eastell R, 39. Ross PD, Davis JW, Epstein RS, Was- 47. Storm T, Thamsborg G, Steiniche T,
McClung MR, Papapoulos S, Rizzoli nich RD (1991) Pre-existing fractures Genant HK, Sørenson OH (1990) Ef-
R, Seeman E, Wasnich RD (1999) Di- and bone mass predict vertebral frac- fect of intermittent cyclical etidronate
agnosis and management of osteoporo- ture incidence in women. Ann Intern therapy on bone mass and fracture rate
sis in postmenopausal women: clinical Med 114:919–923 in women with postmenopausal osteo-
guidelines. International Committee for 40. Ross PD, Davis JW, Epstein RS, Was- porosis. N Engl J Med 322:1265–1271
Osteoporosis Clinical Guidelines. Clin nich RD (1992) Ability of vertebral di- 48. Takada M, Wu CY, Lang TF, Genant
Ther 21:1025–44. Review mensions from a single radiograph to HK (1998) Vertebral fracture assess-
35. Nevitt MC, Ettinger B, Black DM, identify fractures. Calcif Tissue Int 51: ment using the lateral scoutview of
Stone K, Jamal SA, Ensrud K, Segal 95–99 computed tomography in comparison
M, Genant HK, Cummings SR (1998) 41. Ross PD, Genant HK, Davis JW, with radiographs. Osteoporos Int 8:
The association of radiographically de- Miller PD, Wasnich RD (1993) Pre- 197–203
tected vertebral fractures with back dicting vertebral fracture incidence 49. Van Kuijk C, Genant HK (1995) Radi-
pain and function: a prospective study from prevalent fractures and bone den- ology in osteoporosis. In: Riggs BL,
Ann Intern Med 128:793–800 sity among non-black, osteoporotic Melton LJ (eds) Osteoporosis. Raven,
36. Nevitt MC, Ross PD, Palermo L, women. Osteoporos Int 3:120–126 New York
Musliner T, Genant HK, Thompson 42. Ross PD, Yhee YK, He YF, Davis JW, 50. Watts NB, Harris ST, Genant HK,
DE (1999) Association of prevalent Kamimoto C, Epstein RS et al (1993) Wasnich RD, Miller PD, Jackson RD,
vertebral fractures, bone density, and A new method for vertebral fracture di- Licata AA, Ross P, Woodson GCI,
alendronate treatment with incident agnosis. J Bone Miner Res 8:167–174 Yanover MJ, Mysiw J, Kohse L, Rao
vertebral fractures: effect of number 43. Sener RN, Ripeckyj GT, Otto PM, MB, Steiger P, Richmond B, Chesnut
and spinal location of fractures. The Rauch RA, Jinkins JR (1993) Recogni- CHI (1990) Intermittent cyclical
Fracture Intervention Trial Research tion of abnormalities on computed etidronate treatment of postmenopausal
Group. Bone 25:613–619 scout images in CT examinations of osteoporosis. N Engl J Med 323:73–79
37. Rea JA, Li J, Blake GM, Steiger P, the head and spine. Neuroradiology 35: 51. World Health Organization (1994) As-
Genant HK, Fogelman I (2000) Visual 229–231 sessment of fracture risk and its appli-
assessment of vertebral deformity by 44. Smith RW, Eyler WR, Mellinger RC cation to screening for postmenopausal
X-ray absorptiometry: a highly predic- (1960) On the incidence of osteoporo- osteoporosis. Report of a WHO study
tive method to exclude vertebral defor- sis. Ann Intern Med 52:773–781 group. WHO technical report series,
mity. Osteoporos Int 11:660–668 45. Spector TD, McCloskey EV, Doyle Geneva, pp 843:1–129
38. Rea JA, Steiger P, Blake GM, Fogel- DV, Kanis JA (1993) Prevalence of 52. Wu CY, Li J, Jergas M, Genant HK
man I (1998) Optimizing data acquisi- vertebral fracture in women and the re- (1994) Semiquantitative and quantita-
tion and analysis of morphometric lationship with bone density and symp- tive assessment of incident fractures:
X-ray absorptiometry. Osteoporos Int toms: the Chingford Study. J Bone comparison of methods. J Bone Miner
8:177–183 Miner Res 8:817–822 Res 9:S157
46. Steiger P, Cummings SR, Genant HK,
Weiss H (1994) Morphometric X-ray
absorptiometry of the spine: correlation
in vivo with morphometric radiogra-
phy. Study of Osteoporotic Fractures
Research Group. Osteoporos Int 4:
238–244